Broadcom No Longer Sells Itself As A Chip Cycle Story

Broadcom No Longer Sells Itself As A Chip Cycle Story

The business management once framed around a cyclical bottom is still growing, but it now gets a paragraph while artificial intelligence gets the company.

Broadcom (AVGO) has returned about 31% over the past year, ahead of the roughly 18% return on the S&P 500, and no one needs to guess why. The more useful question is what got left behind. Set how management framed this business in late 2024 against how it frames it now, and the company you own has changed shape.

The Cyclical Bottom Broadcom Used To Talk About Openly

Back then, the semiconductor cycle was the lead, not a footnote. In September 2024, the framing was blunt: "non-AI semiconductor revenue has stabilized." A quarter later, management put a figure on the trough, a cyclical bottom in non-AI semiconductor revenue of $17.8 billion for fiscal 2024, the kind of full-year milestone that only matters if the cycle is the story. The other half of that pitch was execution on the VMware integration, which management called largely complete in December 2024. This was a diversified operator asking you to watch two reporting segments.

AI Is Now Two And A Half Times The Rest Of The Chip Business

In fiscal Q2 2026, reported in early June, AI semiconductor revenue was a record $10.8 billion, up 143% from a year earlier. Non-AI semiconductors, the franchise the old story ran on, came in at $4.2 billion, up 6%. Same segment, same quarter, same basis: AI is roughly two and a half times everything else the company sells in silicon, and on its own accounting,

AI semiconductors alone were 49% of total company revenue. The gap widens from here, because management guides fiscal Q3 AI semiconductor revenue to $16 billion and non-AI to about $4.5 billion, and expects $56 billion of AI semiconductor revenue in fiscal 2026. The vocabulary moved with the money. What gets led with now is AI bookings of more than $30 billion in a single quarter against $10.8 billion shipped, and a new platform with Apollo and Blackstone to deploy more than 20 gigawatts of compute capacity.

The Non-AI Business Booked More Than It Shipped In Fiscal Q2

The quiet side is not the weak side. Non-AI semiconductor bookings topped $6 billion in fiscal Q2 2026 against the $4.2 billion actually shipped, which management reads as the path to a full cyclical recovery, and the fiscal Q3 guide of about $4.5 billion is up 12% year over year, double the 6% growth just reported. Company-wide, revenue grew 32% over the trailing twelve months, faster than its 29% three-year average pace; net margin of 39% is well clear of its 28% three-year average; operating margin of 44% sits above its 38% average. Broadcom stopped leading with the cycle, even as management still tracks it closely, because the cycle stopped deciding the year.

The Bet Changed, And The Non-AI Line In September Settles It

The verdict splits. On the business, it is mixed: the quiet franchise is booking more than it ships, which points toward recovery, but its actual revenue has only just clawed back to the 2024 bottom. It gets a paragraph instead of the lead simply because AI grew to two and a half times its size. On the identity, it is a real change. If you bought a diversified compounder with an AI kicker, you now hold a concentrated bet on a handful of AI compute buyers, underwritten by management's fiscal 2027 guide of more than $100 billion in AI semiconductor revenue, and that changes what you are actually paying for.

The line that settles it is non-AI semiconductor revenue when fiscal Q3 is reported on September 2, 2026: clear the $4.5 billion guide and the old franchise has at least clawed back to the level it bottomed at two years ago - real, but not compounding. Fall short, and it hasn't even recovered that far, and the AI story stops being one leg of this stock and becomes nearly all of it. If you are underwriting a number that has to keep climbing, tracking companies whose guidance keeps rising is the discipline this position asks for.

When The Business You Bought Becomes A Different Business

Broadcom did nothing wrong by moving. Businesses follow demand, and this one followed it while its margins improved rather than eroded. The catch for a holder is that your concentration changed without you making a single decision: the same shares now carry far more AI exposure than they did in late 2024. That is the argument for a system rather than a conviction. The Trefis High Quality Portfolio re-balances on rules instead of headlines, which keeps a position sized to your plan when the business inside it becomes a different bet. The Trefis High Quality (HQ) Portfolio has a track record of outpacing a benchmark that combines the three major indices - the S&P 500, S&P Mid-cap, and Russell 2000.